V Learning Companion: How to read the Visa research

The V research family (page counts are those of the approved PDF editions, which this site does not host):

• Deep Dive: A great network, and a price that already assumes it stays one (50 pages). Public preview; the Deep Dive page says how to read the full report.
• Research Note: Priced as if the network cannot be dented (8 pages).
• Technical Appendix: The model's annual schedules (8 pages).
• Learning Companion: How to read the Visa research (14 pages).

---

This edition is for a reader who wants to understand how the Visa judgment was reached, not only what it is. It uses Visa's own numbers as the worked examples, defines each term the first time it matters, and points out where the arithmetic ends and judgment begins. It teaches how the reasoning works; it explains the rating without arguing for it, and it is not advice.

1. What Visa sells, and why that is unusual

Most companies make something and sell it. Visa runs a network. When you pay with a Visa card, four parties are involved: you (the cardholder), your bank (the issuer), the merchant, and the merchant's bank. Visa sits in the middle and earns fees for moving the information and the settlement between them. It generally does not lend to cardholders, and although it sets the default rate for interchange, the fee that merchants' banks pay issuers on each purchase when no other terms have been agreed, it generally does not receive interchange: that fee is the issuers' compensation. Four charges are worth keeping apart. Interchange goes from the merchant's bank to the issuer; Visa's own network fees, paid by issuers and acquirers, are what Visa reports as revenue; the rates and fees a cardholder pays are set by the issuer, not by Visa; and the merchant discount rate is what the merchant's bank charges the merchant. Visa's annual report for fiscal 2026 says the fees Visa receives are not derived from interchange or the merchant discount rate. Because it takes little credit risk and needs little capital, the part of its revenue that survives costs is unusually large. That is the root of its 67.5% adjusted operating margin.

The first lesson is that a high margin is a fact about a business model, and not by itself a reason to buy the shares. A high margin tells you the business is good. Whether the shares are attractive depends on what you pay for that goodness and how long it lasts, and those are different questions.

2. From gross revenue to net revenue

Visa reports four kinds of revenue. In fiscal 2026 (the year ended 09-30-2026) they were:

Check it yourself, in billions of dollars as the Deep Dive's revenue table prints them: 19.918 + 23.374 + 15.251 + 5.677 = 64.220, and 64.220 less 18.317 is 45.903. The incentive ratio, 18.317 divided by 64.220, is 28.5%. Because incentives are deducted from revenue instead of appearing as a cost, they are easy to overlook, yet they are Visa's largest recurring way of paying to protect its position. A rising ratio means clients are keeping a larger share of what flows through the network, and from fiscal 2023 to fiscal 2026 it rose from 27.4% to 28.5%.

A note on method that matters later: in our model, net revenue is built as the sum of the four lines less incentives. That is a modeling identity we chose because it matches how the company reports, and it is not an economic law. Real incentives are negotiated contract by contract; the ratio is our simplification.

3. Volume, yield and why revenue can grow faster than activity

Each revenue line can be split into a volume driver and a yield. Volume is how much flows through (dollars spent, transactions, cross-border dollars); yield is the revenue Visa earns per unit. In fiscal 2026 payments volume and transactions each grew about 9% in constant dollars, yet data processing revenue grew 16.9%. The difference is yield: Visa earned more per transaction through pricing changes and through value-added services such as consulting, risk and fraud products. International transaction revenue went the other way, growing 7.7% while cross-border volume grew about 15% in nominal terms, which the company attributes to lower currency volatility, something Visa earns from, and to business mix.

This is the first place judgment enters. Volume is relatively easy to project because it follows consumer spending. Yield is not, because part of it is price, which can be reversed by competition or regulation, and part is new products, which may or may not keep selling. The company has not said how much of its services growth is price. The staff of the Securities and Exchange Commission asked it to quantify that, and the annual report filed 11-05-2026 lists pricing among the drivers without giving a figure for it. Until the company does, any forecast has to carry yield as an assumption and show what happens if the assumption is wrong. Ours assumes yield keeps adding to growth but at half to a quarter of the fiscal 2026 rate.

4. Reported earnings and adjusted earnings

Visa reports two kinds of operating profit. The GAAP figure, $28.1 billion or a 61.2% margin, follows accounting rules. The adjusted figure removes items management considers unusual: litigation, severance, amortization of acquired intangibles, acquisition costs and a charitable donation. That raises the margin to 67.5%. Adjusted earnings per share were $13.27 against GAAP earnings per share of $12.32, and litigation accounts for most of the gap.

How should you treat the difference? Not by choosing one and ignoring the other. An adjustment is a claim that an item will not recur. Visa has recorded a litigation provision in each of the last five fiscal years, the two largest $2.6 billion in fiscal 2025 and $1.8 billion in fiscal 2026, so it has recurred; to treat it as zero would flatter the business. Yet much of it is paid from an escrow that is funded by the class B shareholders, so charging all of it to class A holders would overstate their burden. The research therefore uses the adjusted cost base (which keeps share-based pay as a cost) and charges a recurring allowance of 1.0% of net revenue for ordinary legal costs that fall on class A, an assumption anchored on fiscal 2026 and not an estimate for the Department of Justice debit case, which appears only in the downside's hypothetical charge. The lesson is general: when a company adjusts its own numbers, ask which adjusted items you believe will vanish, and which you will put back.

5. Cash flow, and the difference between earnings and cash

Free cash flow here means operating cash flow less capital expenditure. In fiscal 2026 that was $23.7 billion less $1.7 billion, or $22.1 billion, equal to 94% of GAAP net income. For a company with few physical assets this is typical: earnings convert to cash almost completely.

Notice what the company did with it. It paid $20.5 billion in repurchases and $5.1 billion in dividends, $25.6 billion in all, or 116% of free cash flow. The difference came out of the balance sheet. That is allowed, and it supports the share price while it lasts, but it cannot continue forever, and it tells you not to treat a buyback as a free source of per-share growth. The research assumes repurchases at the fiscal 2026 pace only for the 32 days between the balance sheet date and the price date.

6. How many shares? The as-converted count

Valuing a company per share requires knowing how many shares exist, and Visa has several classes. Class A is what you can trade. Class B and class C shares, and some preferred stock, convert into class A at fixed rates, so an owner of class A is effectively sharing future profits with them. The sensible count is therefore as-converted: class A plus everything that converts, multiplied by its rate. At 09-30-2026 it was 1,728 class A + (4 × 9 class C) + (2 × 1.5400 class B-1) + 1 class B-2 as converted + (61 × 1.4773 class B-3) + 10 preferred as converted, which comes to 1,868.2 million. The counts are rounded to millions in the filings, so the result is uncertain by roughly three million shares. We then reduce it by 4.5 million for assumed repurchases in the 32 days to the price date, giving 1,863.7 million. As a check, the annual report filed 11-05-2026 gives the class counts at 10-29-2026, and at the same rates they convert to 1,864.4 million, within 0.04% of that figure.

7. Why litigation does not simply reduce the value

Visa has paid billions into an escrow to cover claims from merchants about interchange. Under the retrospective responsibility plan, each deposit shrinks the number of class A shares into which class B converts, at the price on the day. In effect the class B holders bear the cost through dilution, not the class A holders. On 09-18-2026 a $405 million deposit reduced the class B conversion rates; the class B-2 rate, for example, went from 1.5014 class A shares at 06-30-2026 to 1.4924 after the deposit.

The consequence for valuation is subtle and easy to get wrong. If you subtracted the litigation cash from Visa's value and also used the larger share count at the price date, you would count the same cost twice. The model avoids that by charging no covered-litigation cash and keeping that count. The shortcut is valid only while there are enough class B shares to absorb the claims: they are worth $40.1 billion at the 11-02-2026 close against $17.4 billion of interchange at issue. The $17.4 billion is a measure of exposure, not an estimate of what will be paid, and the annual report puts the same measure at $15.9 billion for 10-01-2026. Beyond the buffer, holders must reimburse deposits in cash, and the shortcut stops working.

8. Turning a business into a value: discounting

A discounted cash flow values a company at today's worth of all the cash it will produce for its owners. Because money later is worth less than money now, each future amount is divided by one plus a discount rate for every year of waiting. At a rate of 9.06%, a payment of 100 five years from now is worth about 65 today.

The rate has to reflect risk. We build it, as most practitioners do, from the return on a risk-free asset (the ten-year Treasury yield, here the official par yield of 4.91% for the price date), plus a premium for owning a share of a business, scaled by the stock's sensitivity to the market, called beta. For Visa: 4.91% + 0.90 × 4.75% = 9.19%, the cost of equity. Visa has little debt, only 2.7% of its capital, so the weighted cost of capital is close to the equity cost, at 9.06%.

Be honest about what is in that formula. The risk-free rate is observed. Beta and the equity risk premium are not; they are estimates, and reasonable people use different ones. Ours are judgments, and they sit above some of what we have seen quoted but not registered as evidence, which is why the research shows the value at other settings and does not present one number as the answer.

9. The terminal value, and why it dominates

We forecast cash flows for five years, to fiscal 2031, because no one can see further with any precision. Everything after is bundled into a terminal value: the last year's free cash flow, 39,655 million in the reference, grown by the terminal rate of 4.0% and divided by the gap between the discount rate and that growth: 39,655 × 1.04 ÷ (0.0906 − 0.04) is about 815,378 million when the rate is carried unrounded. That figure is a value at the middle of fiscal 2031, because the flows it stands for arrive in the middle of each later year, and it is discounted to the price date from that point, the same way the five forecast years are. It is 80.9% of the total enterprise value.

That share is the second lesson. When four-fifths of a value comes from the years after the forecast, the value is mainly an opinion about how long and how strongly a business keeps growing. Two things in the reference deserve scrutiny for that reason: growth drops from 8.4% to 4.0% in a single year after fiscal 2031, which is abrupt and which nothing we hold supports or refutes, and the discount rate is higher than some outside readings. Each can be changed, and the effect is large.

10. Reading the equity bridge

Enterprise value is what the operating business is worth. Shareholders own the equity: enterprise value plus cash, less debt and other claims. For the reference case, in millions of dollars: enterprise value of 687,992, plus cash and investments of 14,177 and 2,381 of cash earned between the balance sheet and the price date, less 1,793 of assumed repurchases, less 22,071 of debt, less 362 of accrued litigation net of the escrow that funds it, equals equity of 680,324. Divided by 1,863.7 million as-converted shares, that is $365.04 per share. Each step is tied to a figure in the filings or to a stated assumption, and a reader can rebuild the sum.

11. Scenarios are conditional, not predictions

The research shows three cases: downside $237.26, reference $365.04, upside $462.33. They are not a forecast with a probability on each, and the middle one is not an average of the others. The reference is an illustrative reference case: one fully specified set of assumptions that the other readings are measured against, neither our estimate of what the shares are worth nor a floor. The downside is an enduring impairment of Visa's earning power (a lower-value network that is still highly profitable) and not a temporary recession. The model has no recession year, so a cyclical dip with a recovery has not been valued; we would expect it to be worth more than a permanent impairment, but that is a judgment and not a result. The research states which kind of downside it is because the two are different mistakes: confusing a trough with a permanent loss is how people over- or under-estimate risk.

12. Checking your own work with a second method

A good analyst tries to prove the first answer wrong with a method that does not share its assumptions. Two further methods appear in the research, and only one of them meets that test. The exit lens values the company by what its fiscal 2031 earnings might sell for, applying each multiple to its own case's forecast: 17 times on the downside gives $237.30, 24 times on the reference $406.27 and 30 times on the upside $564.73. Because it uses the same forecasts, it cannot catch a forecast error. The owner-earnings lens starts from fiscal 2026 owner earnings of $24.9 billion, grows them on a single fading path, and reinvests according to the return on new capital. It gave $332.72 at the model's discount rate.

When two methods disagree, the right response is to ask why, not to average them. The averages hide the disagreement, and the disagreement is information.

13. The reverse question

Instead of asking what the shares are worth, ask what would have to be true for them to be worth the price. Holding everything else at the reference, the model matches $425.51 at a cost of capital of 8.34%, or terminal growth of 4.82%, or growth running 3.98 percentage points a year faster than the reference in every forecast year. This is a conditional puzzle, not a statement about what other investors think. It is useful because it shows which assumption the price is most sensitive to. Here that is the discount rate, which is also the least well sourced input.

Keep the dates apart when you read a price. The $425.51 is the close of 11-02-2026, the date of the valuation. The latest settled close, the November 9, 2026 regular-session closing price, was $422.16, which is 0.8% lower, and the reference case of $365.04 sits 13.5% below that close against 14.2% below the earlier one. At that close the cost of capital that would make the reference match the price is 8.38%, against the 8.34% above. A settled close is the price at the end of one regular session, not a live quote. The values were not recalculated for the later date, which is why a refresh of this kind shows both closes side by side and never blends them into one.

14. From a value to a rating

The Blockhaven Rating is a five-level judgment, from 5 COMPELLING to 1 POOR, of risk and reward at the stated price over the stated horizon. It is not a price target, not a buy or sell instruction, and not a calculation: the same numbers can be weighed differently by different analysts. For Visa the level is 3, BALANCED, and how it was reached is the worked example. The facts that weigh against it come first: at the model's own rate every central reading (the reference, the 24-times exit check and the owner-earnings reading) sits below the price, and the downside is 44.2% below the price against an upside only 8.7% above it. The upside case is above the price, but it is a conditional case that bounds the range, not a central reading. What the level rests on is the make-up of the gap. It comes mostly from two inputs that no source we hold settles, the discount rate and the growth path after fiscal 2031, and relaxing both together reaches the price. How much weight unsourced inputs deserve is a judgment, and here it was made so that the price broadly reflects the opportunity and the risks; the reference value itself still sits below the price. The closest competing reading is 2 of 5, which gives those inputs no benefit of the doubt. Keep three things apart: the value the model computes at the valuation date, the rating that compares the price with it, and how the shares trade over the next year, which the research does not forecast. The rating also depends on the price, so it can change when the price moves even if the business and the model do not.

Notice what the report does not print: a grade for how sure we are. Instead it shows the uncertainty itself. The range of values, the sensitivities, the strongest competing reading (a 2 of 5), and the evidence that would change our view are all in the Deep Dive. A single word like "low" or "high" would hide more than it reveals.

15. Reading eight years of history

A single year can mislead, so the Deep Dive sets out eight. Two habits make a long table useful. The first is to look at the line that moves most, which for Visa is international transaction revenue: it fell 19.3% in fiscal 2020 when travel stopped and rose 50.3% in fiscal 2022 when it reopened, so a reader who judged the network by either year alone would be wrong in both directions. The second is to add back a one-off and see what remains. In fiscal 2025 operating income was $24.0 billion and the litigation provision was $2.56 billion. Reported operating margin was 60.0% of net revenue of $40.0 billion, and adding the provision back gives 66.4%. The gap is not an error in either number. It is the difference between what the accounts recorded and what the business earned in a year without the charge, and the Deep Dive uses both views and says which is which.

16. Revenue per dollar and per transaction

Dividing revenue by the volume that drives it gives a yield, and comparing yields across years separates growth in activity from growth in what is charged. In fiscal 2026 Visa's gross revenue was $64.2 billion on nominal payments volume of $15,338 billion, which is 0.419%, or 41.9 cents for each 100 dollars of volume. Client incentives of $18.3 billion are 0.119%, or 11.9 cents, and what remains, 0.299%, is net revenue. In fiscal 2019 the same three figures were 0.338%, 0.072% and 0.267%. Gross yield rose by about 0.08 percentage points and incentives took about 0.05 of them, so the network kept only about 0.03. The same exercise per transaction gives 16.3 cents of net revenue in fiscal 2026 and 16.6 cents in fiscal 2019. A yield is a ratio of two reported numbers, so it needs the caveat that the periods can differ: Visa's revenue year ends 09-30 and its volume year ends 06-30.

17. Incentives as a form of capital

Visa pays clients, mainly the banks that issue its cards, to win and keep their business, and it shows those payments as a deduction from revenue and not as an expense. The company says it generally records upfront and fixed payments as an asset when paid and spreads them against revenue over the contract term, and that variable incentives are recorded as earned. The economic reading is that incentives behave like an investment: they are paid first and earn volume later. In fiscal 2026 they were $18.3 billion against capital expenditure of $1.67 billion, about eleven times as much. The upfront payments do sit on the balance sheet until they are amortized, but the research's free cash flow treats incentives as operating spending and not as capital expenditure. That is why the return on new capital that the model's terminal value implies, which counts only plant and technology, comes out at 461%, a figure that is not meaningful as a hurdle, and why the Deep Dive also reports a return on total invested capital, 56.5%, that includes goodwill and intangibles from acquisitions and is computed on the research's own after-tax operating income, with its assumed tax rate and legal cost.

18. What a peer comparison shows

Comparing Visa with Mastercard and American Express is useful only when the business models are held in view. Mastercard's operating margin for calendar 2025 was 57.6%, against Visa's GAAP margin of 60.0% for fiscal 2025, but Visa's figure includes a large litigation provision, so the comparison depends on which Visa margin is used. American Express issues cards and lends to cardholders, so its net income was 15.0% of its revenues net of interest expense and its card member rewards were 49.2% of its discount revenue, which is a different economic model and not a weaker operator. The lesson is to compare like with like, to say when two companies' fiscal calendars differ, and to treat one competitor's multiple as context and not as a valuation.

19. Payout, free cash flow and the share count

A company can return more cash than it earns, for a while, by drawing on its balance sheet. In fiscal 2026 Visa's repurchases were $20.5 billion and its dividends $5.11 billion, together $25.6 billion, against free cash flow of $22.1 billion, which is a payout of 115.9%. The shortfall came from cash. Repurchases also explain why earnings per share grows faster than net income: over the eight years diluted shares fell 16.1%, so diluted earnings per share compounded at 12.7% a year against 10.0% for net income. A buyback adds to per-share value only if the shares are repurchased for less than they are worth, and on our reference value of $365.04 a repurchase at the $425.51 close retires less value than it costs.

20. Growth needs reinvestment

A terminal value assumes the business grows forever, and growth normally needs capital. In the reference, growth of 4.0% after fiscal 2031 comes with reinvestment of 0.87% of after-tax operating income, which implies a return on each new dollar of 461%. If instead each new dollar earned only Visa's fiscal 2026 return on total capital, 56.5% on the research's own after-tax operating income, the same 4.0% growth would need 4.0 divided by 56.5, or 7.08%, of after-tax operating income to be reinvested, and the reference would be $346.35 and not $365.04. Neither number is the answer. The first overstates what an acquired dollar earns and the second understates what an organic one does, so the honest course is to show both. With everything else held, this one test moves the value between $346.35 and $365.04; that is the size of the question, not a range for the value, which other assumptions move further.

21. A bull case and a bear case

Two cases that argue opposite directions are only useful if each is the strongest honest version of its side. The Deep Dive writes each one the same way: what it believes, the figures it rests on, what it has to assume that the record does not show, and what the model is worth under it. The bull case rests on a record in which net revenue grew less than 10.0% only in the pandemic year, over the seven growth years observed, and is worth $462.33 per share in the upside case. The bear case rests on rising client incentives and payouts above free cash flow and is worth $237.26 in the downside case. The reader's job is to ask of each which of its assumptions the record already supports, and which it only hopes for.

Short glossary

• As-converted shares: the number of class A shares that all share classes would equal if converted at their stated rates.
• Beta: how strongly a share tends to move with the market.
• Client incentives: payments Visa makes to banks and others, deducted from revenue.
• Enterprise value: the value of the operating business before cash and debt.
• Equity risk premium: the extra return investors require for owning shares rather than a risk-free asset.
• Free cash flow: operating cash flow less capital expenditure.
• Interchange: the fee a merchant's bank pays the cardholder's bank on each purchase. Visa sets the default rate, which applies when no other terms have been agreed, but the fee generally goes to the issuer and not to Visa.
• Reverse valuation: finding the assumption that would make the model equal the price, holding the rest fixed.
• Terminal value: the value of all cash flows after the explicit forecast.
• WACC (weighted average cost of capital): the blended return required by owners of both debt and equity.

What this edition does not cover

It does not give the full evidence, the sources by locator or the disclosures; the Deep Dive does. The fiscal 2026 net revenue, net income and earnings per share it uses agree with the annual report filed 11-05-2026.

Assumptions and disclosure

This edition teaches how the research reasons; the evidence, the sources and the full disclosures are in the Deep Dive. The fiscal 2026 figures it uses agree with the annual report filed 11-05-2026. This is research, not personalized advice.

Important disclosures

Blockhaven research is provided for general informational and educational purposes. It is not individualized investment advice and does not take into account any reader's financial circumstances, investment objectives, risk tolerance, tax situation or portfolio.

Nothing in this report is an offer or a solicitation to buy or sell a security or other financial instrument. Reading or accessing Blockhaven research does not by itself create an investment-advisory, fiduciary, brokerage or client relationship.

Investing involves risk, including the possible loss of principal. Scenario values, terminal values, valuation ranges, sensitivities and forward-looking assumptions are analytical tools, not guarantees or promises of future performance, and actual results may differ materially.

Blockhaven uses information from company filings, company publications, market-data providers and other sources it believes to be reliable, but that information may be incomplete, delayed, revised or incorrect.

Opinions, ratings and interpretations reflect Blockhaven's research judgment as of the report date and may change as new evidence becomes available.

Positions and conflicts

As of November 6, 2026, Blockhaven's owner held no shares, options, or other direct derivatives in Visa Inc.; had no material positions in Visa Inc. through entities the owner controls; and had no paid financial relationships with Visa Inc.

© 2026 Blockhaven LLC. All rights reserved.